Search

Supplementary Table S1 from Gene expression profiling-based identification of cell-surface targets for developing multimeric ligands in pancreatic cancer

Multimeric ligands are ligands that contain multiple binding domains that simultaneously target multiple cell-surface proteins. Due to cooperative binding, multimeric ligands can have high avidity for cells (tumor) expressing all targeting proteins and only show minimal binding to cells (normal tissues) expressing none or only some of the targets. Identifying combinations of targets that concurrently express in tumor cells but not in normal cells is a challenging task. Here, we describe a novel approach for identifying such combinations using genome-wide gene expression profiling followed by immunohistochemistry. We first generated a database of mRNA gene expression profiles for 28 pancreatic cancer specimens and 103 normal tissue samples representing 28 unique tissue/cell types using DNA microarrays. The expression data for genes that encode proteins with cell-surface epitopes were then extracted from the database and analyzed using a novel multivariate rule-based computational approach to identify gene combinations that are expressed at an efficient binding level in tumors but not in normal tissues. These combinations were further ranked according to the proportion of tumor samples that expressed the sets at efficient levels. Protein expression of the genes contained in the top ranked combinations was confirmed using immunohistochemistry on a pancreatic tumor tissue and normal tissue microarrays. Coexpression of targets was further validated by their combined expression in pancreatic cancer cell lines using immunocytochemistry. These validated gene combinations thus encompass a list of cell-surface targets that can be used to develop multimeric ligands for the imaging and treatment of pancreatic cancer. [Mol Cancer Ther 2008;7(9):3071–80]

Two issues are critical to the development of effective cancer-drug combinations. First, it is necessary to determine common combinations of alterations that exert strong control over proliferation and survival regulation for the general type of cancer being considered. Second, it is necessary to have a drug testing method that allows one to assess the variety of responses that can be provoked by drugs acting at key points in the cellular processes dictating proliferation and survival. Utilizing a previously reported GFP (green fluorescent protein) reporter-based technology that provides dynamic measurements of individual reporters in individual cells, the present paper proposes a dynamical systems approach to these issues. It involves a three-state experimental design: (1) formulate an oncologic pathway model of relevant processes; (2) perturb the pathways with the test drug and drugs with known effects on components of the pathways of interest; and (3) measure process activity indicators at various points on cell populations. This design addresses the fundamental problems in the design and analysis of combinatorial drug treatments. We apply the dynamical approach to three issues in the context of colon cancer cell lines: (1) identification of cell subpopulations possessing differing degrees of drug sensitivity; (2) the consequences of different drug dosing strategies on cellular processes; and (3) assessing the consequences of combinatorial versus monotherapy. Finally, we illustrate how the dynamical systems approach leads to a mechanistic hypothesis in the colon cancer HCT116 cell line.

The inhibitory effect of caveolin on the cellular response to growth factor stimulation is well established. Given the significant overlap in signaling pathways involved in regulating cell proliferation and stress responsiveness, we hypothesized that caveolin would also affect a cell's ability to respond to environmental stress. Here we investigated the ability of caveolin-1 to modulate the cellular response to sodium arsenite and thereby alter survival of the human cell lines 293 and HeLa. Cells stably transfected with caveolin-1 were found to be much more sensitive to the toxic effects of sodium arsenite than either untransfected parental cells or parental cells transfected with an empty vector. Unexpectedly, the caveolin-overexpressing cells also exhibited a significant activation of the phosphatidylinositol 3-kinase (PI3K)/Akt pathway, which additional studies suggested was likely due to decreased neutral sphingomyelinase activity and ceramide synthesis. In contrast to its extensively documented antiapoptotic influence, the elevated activity of Akt appears to be important in sensitizing caveolin-expressing cells to arsenite-induced toxicity, as both pretreatment of cells with the PI3K inhibitor wortmannin and overexpression of a dominant-negative Akt mutant markedly improved the survival of arsenite-treated cells. This death-promoting influence of the PI3K/Akt pathway in caveolin-overexpressing cells appeared not to be unique to sodium arsenite, as wortmannin pretreatment also resulted in increased survival in the presence of H(2)O(2). In summary, our results indicate that caveolin-induced upregulation of the PI3K/Akt signaling pathway, which appears to be a death signal in the presence of arsenite and H(2)O(2), sensitizes cells to environmental stress.

All cells depend on sterols and isoprenoids derived from mevalonate (MVA) for growth, differentiation, and maintenance of homeostatic functions. In plants, environmental insults like heat and sunlight trigger the synthesis of isoprene, also derived from MVA, and this phenomenon has been associated with enhanced tolerance to heat. Here, we show that in human prostate adenocarcinoma PC-3M cells heat shock leads to activation of the MVA pathway. This is characterized by a dose- and time-dependent elevation in 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGR) activity, enhanced sterol and isoprenoid synthesis, and increased protein prenylation. Furthermore, prenylation and subsequent membrane localization of Ras, a central player in cell signaling, was rapidly induced following heat stress. These effects were dose-dependent, augmented with repeated insults, and were prevented by culturing cells in the presence of lovastatin, a competitive inhibitor of HMGR. Enhanced Ras maturation by heat stress was also associated with a heightened activation of extracellular signal-regulated kinase (ERK), a key mediator of both mitogenic and stress signaling pathways, in response to subsequent growth factor stimulation. Thus, activation of the MVA pathway may constitute an important adaptive host response to stress, and have significant implications to carcinogenesis.

The aromatic fatty acid phenylacetate, a common metabolite of phenylalanine, shows promise as a relatively non-toxic drug for cancer treatment. This slowly metabolized fatty acid alters tumor cell lipid metabolism causing, among other effects, inhibition of protein prenylation critical to malignant growth. In pursuit of more potent analogues, we have examined the activity of related compounds against tumor cell lines established from patients with advanced prostatic carcinoma, glioblastomas, and malignant melanoma. Like phenylacetate, derivatives containing α-carbon or ring substitutions induced cytostasis and phenotypic reversion at non-toxic concentrations. Potency was correlated with the degree of calculated lipophilicity of the aromatic fatty acid, and the extent of inhibition of protein prenylation. Remarkably, a parallel cytostatic activity was reported in embryonic plant cells, which respond to phenylacetate and its analogues in the same concentration range and the same rank order of lipophilicity. These data suggest that phenylacetate and its analogues may act through common mechanisms to inhibit the growth of vastly divergent, undifferentiated cell types, and provide a basis for the development of new agents for the treatment of human malignancies.

High-content cell imaging based on fluorescent protein reporters has recently been used to track the transcriptional activities of multiple genes under different external stimuli for extended periods. This technology enhances our ability to discover treatment-induced regulatory mechanisms, temporally order their onsets and recognize their relationships. To fully realize these possibilities and explore their potential in biological and pharmaceutical applications, we introduce a new data processing procedure to extract information about the dynamics of cell processes based on this technology. The proposed procedure contains two parts: (1) image processing, where the fluorescent images are processed to identify individual cells and allow their transcriptional activity levels to be quantified; and (2) data representation, where the extracted time course data are summarized and represented in a way that facilitates efficient evaluation. Experiments show that the proposed procedure achieves fast and robust image segmentation with sufficient accuracy. The extracted cellular dynamics are highly reproducible and sensitive enough to detect subtle activity differences and identify mechanisms responding to selected perturbations. This method should be able to help biologists identify the alterations of cellular mechanisms that allow drug candidates to change cell behavior and thereby improve the efficiency of drug discovery and treatment design.

Since its origin in the mid-1990s, gene expression profiling by microarray has become a productive and useful tool in basic science and preclinical research. Current dual-color, high-density cDNA oligo arrays contain 60-mer detectors for the whole human genome. With this powerful technology, expression of RNA samples from cell lines or tissue can be assessed, revealing specific gene expression signatures. The technique includes three major steps: (1) isolation and purification of RNA from cells or tissues, (2) labeling of total RNA, and (3) hybridization with Agilent cDNA microarrays. Conveniently, this technique can be performed with as little as 50 ng of purified total RNA; however, it is important to keep in mind that the quality of the RNA template, namely the level of sample degradation and the presence of contaminants that are carried over from the starting material or introduced during RNA isolation, can significantly impact the efficiency of the labeling reaction and the reliability of the hybridization. In this chapter, the details of each step of this technique are explained thoroughly, while highlighting the key issues that can prevent a failed hybridization.

Since its origin in the mid-1990s, gene expression profiling by microarray has become a productive and useful tool in basic science and preclinical research. Current dual-color, high-density cDNA oligo arrays contain 60-mer detectors for the whole human genome. With this powerful technology, expression of RNA samples from cell lines or tissue can be assessed, revealing specific gene expression signatures. The technique includes three major steps: (1) isolation and purification of RNA from cells or tissues, (2) labeling of total RNA, and (3) hybridization with Agilent cDNA microarrays. Conveniently, this technique can be performed with as little as 50 ng of purified total RNA; however, it is important to keep in mind that the quality of the RNA template, namely the level of sample degradation and the presence of contaminants that are carried over from the starting material or introduced during RNA isolation, can significantly impact the efficiency of the labeling reaction and the reliability of the hybridization. In this chapter, the details of each step of this technique are explained thoroughly, while highlighting the key issues that can prevent a failed hybridization.

Multimeric ligands are ligands that contain multiple binding domains that simultaneously target multiple cell-surface proteins. Due to cooperative binding, multimeric ligands can have high avidity for cells (tumor) expressing all targeting proteins and only show minimal binding to cells (normal tissues) expressing none or only some of the targets. Identifying combinations of targets that concurrently express in tumor cells but not in normal cells is a challenging task. Here, we describe a novel approach for identifying such combinations using genome-wide gene expression profiling followed by immunohistochemistry. We first generated a database of mRNA gene expression profiles for 28 pancreatic cancer specimens and 103 normal tissue samples representing 28 unique tissue/cell types using DNA microarrays. The expression data for genes that encode proteins with cell-surface epitopes were then extracted from the database and analyzed using a novel multivariate rule-based computational approach to identify gene combinations that are expressed at an efficient binding level in tumors but not in normal tissues. These combinations were further ranked according to the proportion of tumor samples that expressed the sets at efficient levels. Protein expression of the genes contained in the top ranked combinations was confirmed using immunohistochemistry on a pancreatic tumor tissue and normal tissue microarrays. Coexpression of targets was further validated by their combined expression in pancreatic cancer cell lines using immunocytochemistry. These validated gene combinations thus encompass a list of cell-surface targets that can be used to develop multimeric ligands for the imaging and treatment of pancreatic cancer. [Mol Cancer Ther 2008;7(9):3071–80]

Oxidative stress is believed to be an important factor in the development of age-related diseases, and studies in lower organisms have established links between oxidative stress tolerance and longevity. We have hypothesized that aging is associated with a reduced ability to mount acute host defenses to oxidant injury, which increases the vulnerability of aged cells to stress. We tested this hypothesis by using primary hepatocytes from young (4-6 months) and aged (24-26 months) rats. Old hepatocytes were more sensitive to H2O2-induced apoptosis than were young cells. Lower survival is associated with reduced activations of extracellular signal-regulated kinase (ERK) and Akt kinase, both of which protect against oxidant injury. That reduced ERK and Akt activities contribute to lower survival of aged cells was supported by additional findings. First, pharmacologic inhibition of ERK and Akt activation in young cells markedly increased their sensitivity to H2O2. Second, caloric restriction, which increases rodent life span and delays the onset of many age-related declines in physiologic function, prevented loss in ERK and Akt activation by H2O2 and enhanced survival of old hepatocytes to levels similar to those of young cells. Strategies aimed at boosting these host responses to acute oxidant injury could have significant anti-aging benefits.

Any attempt to eradicate the heterogeneous cell population of a tumour mass would require the use of appropriate combination treatment protocols. The antitumour effects of interferon alpha (IFN alpha) in combination with AS2-1, the hydrolysis product of 3-phenylacetyl-amino-2,6-piperidinedione, were examined using several human tumour cell lines as a model. These included the malignant melanoma A375, adenocarcinoma of the prostate PC3 (hormone-insensitive bone metastasis), and the erythroleukaemia line K562. AS2-1 suppressed tumour growth through non-toxic mechanisms, with 1 mg/ml causing approximately 50% inhibition of the melanoma and prostate tumour cell proliferation. By contrast, primary normal human skin fibroblasts were significantly less sensitive to the antiproliferative effect of AS2-1. Suppression of tumour growth was seen also with AS2-1 treatment of the erythroleukaemia K562; in these cultures the drug also induced dose-dependent differentiation, as indicated by the increased haemoglobin production. Interestingly, addition of low doses of IFN alpha markedly enhanced the antitumour and differentiating effects observed with AS2-1. Treatment with 200-300 IU/ml of IFN (which caused about 20% inhibition of growth) together with 1 mg/ml of AS2-1 resulted in over 80% inhibition of the melanoma and prostate cancer cell proliferation, suggesting a synergistic activity of the two agents. This was substantiated by quantitative analysis of the differentiation induced in K562 erythroleukaemia. It appears, therefore, that IFN alpha and AS2-1 may act through synergistic mechanisms to effectively inhibit tumour growth and promote differentiation in a variety of human malignant cell lines.

Cinnamic acid, a naturally occurring aromatic fatty acid of low toxicity, has a long history of human exposure. We now show that cinnamic acid induces cytostasis and a reversal of malignant properties of human tumor cells in vitro. The concentration causing a 50% reduction of cell proliferation (IC50) ranged from 1 to 4.5 mM in glioblastoma, melanoma, prostate and lung carcinoma cells. Using melanoma cells as a model, we found that cinnamic acid induces cell differentiation as evidenced by morphological changes and increased melanin production. Moreover, treated cells had reduced invasive capacity associated with modulation of expression of genes implicated in tumor metastasis (collagenase type IV, and tissue inhibitor metalloproteinase 2) and immunogenicity (HLA-A3, class-I major histocompatibility antigen). Further molecular analysis indicated that the anti-tumor activity of cinnamic acid may be due in part to the inhibition of protein isoprenylation known to block mitogenic signal transduction. The results presented here identify cinnamic acid as a new member of the aromatic fatty acid class of differentiation-inducers with potential use in cancer intervention.

Phenylacetate, a natural metabolite of phenylalanine which was originally described as a plant growth hormone, has recently gained attention as a possible differentiation inducer for a variety of human tumor cell types. This interest prompted us to assess the ability of sodium phenylacetate (NaPA) to promote the differentiation of human neuroblastoma cells, both alone and in combination with retinoic acid (RA), a known inducer of neuroblastoma differentiation and maturation. Using the LA-N-5 cell line, we have determined that NaPA can stimulate the differentiation of neuroblastoma cells, as evidenced by dose-dependent inhibition of cell proliferation, neurite outgrowth, increased acetylcholinesterase activity and reduction of N-myc expression. Furthermore, NaPA and RA synergized in inducing differentiation, in that combination treatment resulted in cessation of cell growth along with morphologic and biochemical changes indicative of the loss of malignant properties. We have determined that NaPA can markedly enhance mRNA levels of the nuclear RA receptor-beta (RAR beta) in LA-N-5 cells prior to morphologic or other phenotypic changes induced by this compound. This effect appeared to be distinct from the ability of NaPA to alter tumor cell lipid metabolism via inhibition of protein isoprenylation. Thus among its varied effects on LA-N-5 cells, NaPA appears to interact with the RA pathway at the nuclear level by up-regulating RAR beta expression.

Alterations in the expression of ras oncogenes are characteristic of a wide variety of human neoplasms. Accumulating evidence has linked elevated ras expression with disease progression and with failure of tumors to respond to conventional therapies, including radiotherapy and certain chemotherapies. These observations led us to investigate the response of ras-transformed cells to the differentiation-inducer phenylacetate (PA). Using gene transfer models, we show that PA caused cytostasis in ras-transformed mesenchymal cells, associated with increased expression of 2',5'-oligoadenylate synthetase, an enzyme implicated in negative growth control. PA also induced phenotypic reversion characterized by loss of anchorage-independent growth, reduced invasiveness and increased expression of collagen alpha type I, a marker of cell differentiation. The anti-tumor activity of PA was observed in cases involving either Ha- or Ki-ras and was independent of the mode of oncogene activation. Interestingly, in contrast to their relative resistance to radiation and doxorubicin, ras-transformed cells were significantly more sensitive to PA than their parental cells. The profound changes in tumor cell and molecular biology were associated with reduced isoprenylation of the ras-encoded p21. Our results indicate that PA can suppress the growth of ras-transformed cells, resistant otherwise to free-radical based therapies, through interference with p21ras isoprenylation, critical to signal transduction and maintenance of the malignant phenotype.

The increasing incidence of melanoma and the poor responsiveness of disseminated disease to conventional treatments call for the development of new therapeutic approaches. Phenylacetate, a nontoxic differentiation inducer, can suppress the growth of other neuro- ectodermal tumors, i.e., gliomas, in laboratory models and in humans. This finding led us to explore the efficacy of phenylacetate and related aromatic fatty acids in melanoma. Phenylacetate and phenylbutyrate were found to a) induce selective cytostasis and maturation of cultured human melanoma cells, b) modulate the expression of genes implicated in tumor metastasis (type IV collagenase and tissue inhibitor of metalloproteinases-2) and immunogenicity (HLA class I); and c) enhance the efficacy of other agents of clinical interest, including retinoids, interferon-alpha, suramin, and 5-aza-2'-deoxycytidine. Reflecting on the phenotypic heterogeneity of melanoma, the degree of biologic alterations induced by phenylacetate, phenylbutyrate varied significantly among the tumor cell lines tested. Although losing invasive capacity and tumorigenicity in athymic mice, poorly differentiated cells exhibited only a marginal change in morphology, remained amelanotic, and resumed growth after treatment was discontinued. By contrast, treatment of melanoma cells that were in a more advanced stage of maturation resulted in profound alterations in cell growth, morphology, and pigmentation consistent with terminal differentiation. The in vitro antitumor activity was observed with nontoxic, pharmacologic concentrations of phenylacetate and phenylbutyrate, suggesting potential clinical use of these drugs in the treatment of melanomas.

The analysis of the dynamics of cellular responses to drugs is currently bounded by inherent limitations of the methods most often used to assess levels of promoter activity and protein abundance or activation. The sensitivity of these assays is limited by the averaging of analyte over many cells, which can produce an average that is not accurately descriptive of the wide range of activities of members of the population examined. The ability of these assays to indicate the trajectory of the various responses is limited by a very infrequent sampling rate, frequently 2–4 time points spread over many hours. These limits are further exacerbated by populations of treated cells' inhomogeneity in the timing of their own response to the drugs. To improve the determination of the functional dynamics of cells' responses to drugs, PBS-Bio has developed technology that allows gathering data from living cells using fluorescent reporter technology. This technology allows measurements to be made serially on single populations of cells at the individual cell level, providing both high sensitivity in terms of when even a small fraction of the cells exhibit significant shifts of a particular activity, the rate of shift of cells in the population to this different activity and the ability to temporally order the timing of events relative to each other. Experimentation on two cell lines, A549 and PC3, using a variety of drugs that affect different cellular processes and a set of fluorescent reporters that provide information on a variety of cellular processes it was possible to show that the modified cardenolide drug, UNBS1450 (Unibioscreen), provoked a very early reduction in the expression of the MCL1 and MYC genes, followed by induction of a strong apoptotic response in both cell lines. As loss of MCL1 provoked by other means has been shown to be sufficient to induce apoptosis in both these lines, a very strong case can be made for UNBS1450's ability to provoke Mcl1 protein reduction as the drug's mode of apoptotic induction. This finding would also implicate Mcl1 as a useful biomarker of the possible efficacy of UNBS1450. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr LB-277. doi:10.1158/1538-7445.AM2011-LB-277

e15014 Background: Rectal cancer differs from colon cancer in terms of prognosis. Here we present genome-wide expression analysis of 79 colon and rectal tumors and gene dosage analysis of 34 archival colorectal tumors in paraffin (FFPE). Methods: Expression data were acquired ( http://expo.intgen.org/geo/home.do ) from Affymetrix experiments analyzed with GeneSpring software version 7.3 (Silicon Genetics, CA). Expression data were normalized ’per chip’ normalized to the 50th percentile of all values; and ’per gene’ normalized to the median expression level across all samples. Expression Project for Oncology (expO) samples included colon (n=50) and rectum (n=29). Differentially expressed genes were identified by parametric test for which variances were not assumed equal (Welch ANOVA). Array comparative genomic hybridization (aCGH) were performed separately on an independent set of FFPE colorectal tumors (n=34) by 244 K microarrays for CGH, Agilent Technologies, CA and with platform embedded analyses tools of CGH Analytics software and ADM-1 bioinformatics. Functional assay by knockout of HOXA9 promoter by lentiviral sh-RNAi construct was performed on colon cancer cell lines shown to have increased HOXA9 dosage and expression. Results: The analysis of expO colon and rectal cancers generated 42 genes with significant differential expression, 26 showed an increased expression of > 2 fold in colon versus rectal samples. Transcription factor family HOXA9 was the most highly expressed (3.8 fold) in colon versus rectal cancer. The aCGH data showed a low level gain of the HOXA gene to be the most frequent dosage alteration (38%). Validation by IHC and qRT-PCR showed 80% and 76% concordance, respectively. Functional assay by sh-RNAi in cell lines (compared to vehicle only) showed marked decrease in cellular viability (40 to 60%), marked morphologic change and a significantly increased apoptotic rate. Conclusion: Multiple genome-wide assays have identified HOXA9 as differentially dosed and expressed in colon and rectal tumors. Compelling functional data from sh-RNAi experiments suggests a tumorigenic role for HOXA9 in altered apoptosis. No significant financial relationships to disclose.

3071 An important question that arises when caring for patients referred for phase I clinical trials is, if one studied the patients’ tumors carefully enough, would there be targets in those patients’ tumors for which a therapeutic agent might already be available. Perhaps treating the patient with that therapeutic agent would be better than trying the phase I agent. To address that issue we have performed IHC assays for up to 13 targets (e.g. Her2/neu, ER, c-kit, etc.) as well as a two color Agilent oligonucleotide microarray with 17, 085 unique probes, on tumors from patients with advanced cancer who have exhausted conventional therapy and who were undergoing a procedure for a cancer related matter (e.g. obstruction, ascites, etc.). Ninety-one patients had excess tissue submitted from the procedure, immediately frozen and subjected to microarray analysis and had IHC performed on sections obtained from the paraffin blocks prepared as part of their standard surgical workup. IHC was considered positive if ≥ 30% of tumor cells were scored as +1 or greater staining. For microarray studies the gene expressions were called significantly increased or decreased if the fold change in expression relative to the nominal tissue of origin reference was significant at p≤0.001. For the 91 patients where frozen tumor was submitted 89 (98%) of them had RNA that was of excellent quality. The findings were a surprise in that IHC identified an average of 1.2 targets (for which a conventional therapeutic agent is available) per patient (range 0 - 4 targets). The DMA identified an average of 3.3 targets per patient (range 0 - 8 targets). Overall, a target was found for 89 (98%) of the patients. This data indicates that using IHC and DMA one can consistently find a potential target for which we have a therapeutic agent in patients’ tumors even thought they have progressed on prior therapies. This has implications for patients’ planning on participating in phase I clinical trials. It is important to conduct a prospective clinical trial to determine whether the targets discovered using IHC and DMA can be used to select therapy which will further benefit the patient who would otherwise be participating in a phase I clinical trial. [Table: see text]